## DNAJC4
- **UniProt:** Q9NNZ3 · **batch:** proteostasis-batch-2026-06-07b · **review status:** COMPLETE
- **PN placement:** `Mitochondrial proteostasis|Chaperone|HSP70 system|J-domain containing HSP70 cochaperone` (branch MI) ; **PN-node mapping:** type → mapped/ok_for_propagation_to_go GO:0030544 Hsp70 protein binding (goa_status=more_specific_than_existing_goa); all ancestor nodes no_mapping.
- **Consistency:** Mostly consistent but with a placement nuance and an MF mismatch. Deep research (notes), review and PN agree DNAJC4 is a Tdark, essentially uncharacterized type-III DnaJ/HSP40 co-chaperone (J domain aa 34-99) with a predicted single-pass TM helix; PAN-GO returns 0 evolutionary annotations. No experimental chaperone data exist. Discrepancy 1: PN places DNAJC4 in the **Mitochondrial** branch, but neither the review nor notes give any mitochondrial evidence (review's only locations are membrane/single-pass TM; UniProt gives no mito targeting). Discrepancy 2: PN projects GO:0030544 Hsp70 protein binding as the type MF, but the review's core MF is GO:0051082 unfolded protein binding (chosen deliberately as a cautious family-level call) — GO:0030544 appears nowhere in the review.
- **PN story / NEW pressure:** PN asserts an Hsp70-binding MF (GO:0030544, verified real via OLS) not present in GOA (GOA has only response to unfolded protein, BP) nor in the review. For a J-domain protein this is the canonical co-chaperone MF and is defensible **by family**, but is unverified experimentally for DNAJC4 specifically — consistent with the review's deliberately cautious GO:0051082. Conclusion: a defensible family-level ADD candidate, but **unverified**; do not assert as core.
- **Mapping strategy:** Type-level GO:0030544 is appropriate breadth for a J-domain co-chaperone and is NOT an over-reach (unlike holdase claims). However the Mitochondrial-branch placement is unsupported by any evidence in the dossier and should be checked.
- **Evidence alignment:** PN row carries no reference titles. Review/notes cite PMID:9473517 (cloning paper; UNVERIFIED, not cached), PMID:17500595 (HTT interactome), PMID:32814053 (neurodegeneration interactome). No overlap with PN, but no contradiction.
- **Verdict:** Family-level J-co-chaperone call sound; GO:0030544 is a defensible-but-unverified ADD; mitochondrial placement and the GO:0030544-vs-GO:0051082 MF mismatch need reconciling. **Recommended edits:** [MAP] reconcile Hsp70 protein binding (GO:0030544) projection with the review's GO:0051082 unfolded protein binding — both are family-level/unverified; prefer GO:0030544 only if asserted as predicted, not experimental. [MAP][WB] verify the Mitochondrial-branch assignment — no mitochondrial localization evidence appears in the review or notes.